MSc IB PREREQUISITE COURSE

Management Research Methods

FINAL EXAM PROJECT

Group no.: 2

Active Participants

Name

Exam number

Aleksandra Lozinska

201400231 / 546576

Jim Heg Pedersen

JP93185 / 994790

Rune Kjaer Andersen

RA93262 / 413198

Kasper Lykke Hoffbeck

KH93153 / 413178

Kimie Vestergaard Andersen

KA94025 / 302763

Number of characters excl. blanks, front page, introduction, bibliography and appendix: 72.545

Department of Business Administration

Business and Social Sciences, Aarhus University
December 2014

Page 1 of 67

Table of Content
1 Introduction ....................................................................................................................................... 5
1.1 Background and managerial objectives ..................................................................................... 5
1.1.2 Glossary .............................................................................................................................. 6
1.2 Theoretical framework ............................................................................................................... 6
1.3 Previous research ....................................................................................................................... 6
1.4 Research Question and overview of research design ................................................................. 8
1.5 Hypotheses ................................................................................................................................. 8
1.5.1 National culture: .................................................................................................................. 8
1.5.2 Organizational culture: ........................................................................................................ 9
1.5.3 Language barriers:............................................................................................................... 9
1.5.4 Current employment ......................................................................................................... 10
1.5.5 Social factors: .................................................................................................................... 10
1.5.6 The Danish societal structure ............................................................................................ 10
1.5.7 Career perspective in Denmark ......................................................................................... 11
1.6 Conceptual framework ............................................................................................................. 11
1.6.1 Expats, studying on AU, School of Business and Social Sciences: .................................. 12
1.6.2 Factors that influence the decision to stay or leave: ......................................................... 12
1.6.3 The plus/minus arrow: ...................................................................................................... 13
1.6.4 To stay or leave scale: ....................................................................................................... 13
2 Primary quantitative study .............................................................................................................. 13
2.1 Quantitative study - Method .................................................................................................... 13
2.1.1 Research guide structure: .................................................................................................. 13
2.1.2 Statistical analysis: ............................................................................................................ 14
2.1.3 Choice of scale intervals: .................................................................................................. 14
2.1.4 Sampling technique: .......................................................................................................... 14
2.1.5 Distribution of the questionnaire: ..................................................................................... 15
2.1.6 Pilot testing: ...................................................................................................................... 15
2.2 Factor analysis as a method: .................................................................................................... 15
2.2.1 Exclusion of variables based upon unfulfilled assumptions ............................................. 16
2.2.2 Number of factors ............................................................................................................. 16
2.2.3 Rotation ............................................................................................................................. 17
2.2.4 Factorability of R (after variable/factor reductions) ......................................................... 17
2.2.5 Assumptions for the best fitting model of factors: ............................................................ 18
2.3 Results of factor analysis ......................................................................................................... 20
2.3.1 Discussion of the factors and their connection to the conceptual framework .................. 21
Page 2 of 67

2.4 Multivariate Regression analysis method. ............................................................................... 24

2.4.1 Assumptions ...................................................................................................................... 24
2.4.2 Ratio of Cases to Independent variables ........................................................................... 24
2.4.3 Absence of Outliers among the IVs and on the DV .......................................................... 24
2.4.4 Absence of multicollinearity and singularity .................................................................... 25
2.4.5 Normality, Linearity, and Homoscedasticity and independence of Residuals ................. 25
2.4.6 Absence of Outliers in the Solution .................................................................................. 26
2.4.7 Excluding variables using the enter method ..................................................................... 26
2.5 Results from final multiple regression model .......................................................................... 27
2.5.1 Adjusted R Square ............................................................................................................ 27
2.6

Discussion/perspective ........................................................................................................ 28

2.7 Results ...................................................................................................................................... 28

2.8 Conclusion and modification to the conceptual model ............................................................ 30
3 Study 2 Qualitative.......................................................................................................................... 31
3.1 Method ..................................................................................................................................... 31
3.1.1 Interview structure: ........................................................................................................... 31
3.1.2 Question development: ..................................................................................................... 31
3.1.3 Interview form and choice of interviewee: ....................................................................... 32
3.1.4 Analysis of data:................................................................................................................ 33
3.2 Results ...................................................................................................................................... 34
3.3 Discussion ................................................................................................................................ 37
3.3.1 Modification for the conceptual model ............................................................................. 37
4 General discussion and conclusion ................................................................................................. 37
4.1 Reliability and Validity ............................................................................................................ 37
4.2 Final Conclusion and modifications to the conceptual model ................................................. 40
4.3 Alternative use of mixed method ............................................................................................. 41
Bibliography....................................................................................................................................... 43
Appendix 1) Test of normality of chosen variables within the factor analysis.................................. 45
Appendix 2) Test of linearity between the variables ......................................................................... 46
Appendix 3) SMC values between variables (after exclusion of insignificant variables) ................. 47
Appendix 4) KMO and Bartletts (after exclusion of insignificant variables) .................................. 48
Appendix 5) MSA (with all variables before exclusion) ................................................................... 49
Appendix 6) Rotated component matrix with all variables ............................................................... 50
Appendix 7) Rotated component Matrix without fair_tax ................................................................. 50
Appendix 8) Anti-image matrix (excluded fair_tax) ......................................................................... 51
Appendix 9) Rotated component matrix (excluded fair_tax and dk_individual) .............................. 52
Page 3 of 67

Appendix 10) The best fitting factor matrix ...................................................................................... 52

Appendix 11) The regression model .................................................................................................. 53
Appendix 12) Histogram of the IVs (test of outliers among the IVs) .............................................. 54
Appendix 13) Significance level of all six factors in the regression model ...................................... 55
Appendix 14) Histogram of residuals ................................................................................................ 55
Appendix 15) Residuals - checking for normality, linearity and heteroscedasticity ......................... 56
Appendix 16) Linearity of residuals .................................................................................................. 57
Appendix 17) Histogram of DV......................................................................................................... 57
Appendix 18: Total Variance expected ............................................................................................. 58
Appendix 19: Scree plot ..................................................................................................................... 58
Appendix 20) Communalities for the best fitting model ................................................................... 59
Appendix 21) Transcript of qualitative interview .............................................................................. 60
Codes:............................................................................................................................................. 60
Transcript ....................................................................................................................................... 60
Appendix 22: Survey ......................................................................................................................... 67

Page 4 of 67

1 Introduction
1.1 Background and managerial objectives
In order to compete in the global economy, Denmark needs to fight for international talents which
automatically will attract more international businesses and strengthen the position on the global
market (Regeringen, 2014b). Research from Dansk Industri (2010) suggests that in 2030 the
demand for highly-educated employees in the Danish labor market will exceed the supply, with an
estimated need for at least 30.000 highly-educated employees. Despite the high Danish educational
level; the amount of domestic graduates does not fulfil the demand, which is necessary to attract
Foreign Direct Investments (FDI) (Dansk Industri, 2010). Compared to other countries, Denmark
attracts relatively few FDIs, which is problematic as these investments directly influence economic
growth. Experts points to the highly educated workforce as one of the most significant variables in
the internationalization of multinational companies (Munksgaard, 2014; Regeringen, 2014b).
In 2012 the Danish government spent nearly 1 billion DKK on educating foreigners in Denmark
(Zieler, 2013), which indicates a focus on education of foreign students, in order to qualify them for
the Danish labor market. Denmark is perceived a highly attractive study destination, but studies
show that 47 % of foreign students in Denmark leave the country after graduation, and thereby
complicate the processes of meeting the demand in 2030. Therefore, the investments in education of
foreigners are not optimal in an economic perspective. In order to meet the increasing demand for
highly educated international workforce, the Danish government introduced 17 initiatives in 2014 to
attract and retain international talent. For instance, an initiative is to make Statens
Uddannelsessttte (SU) available for foreign students (Regeringen, 2014b). Moreover, studies
indicate that the GDP effect of retaining an additional fifth of the foreign graduates is 500 million
DKK per year, which makes it highly attractive to retain these students (Dansk Industri, 2012).
The purpose of this research is to identify factors that influence the decision of foreign graduates to
stay in or leave Denmark after graduation. This knowledge can help Danish companies utilize these
factors in the hiring process, and thereby possibly enhance global performance, broaden their
international perspective, and utilize language skills and cultural diversity to increase the
competitive advantage on the global market.

Page 5 of 67

1.1.2 Glossary
An expat is defined as someone who lives outside their native country (Oxford Dictionaries, 2014),
therefore it is assumed, that foreign students and graduates living in Denmark can be defined as
expats.

1.2 Theoretical framework

This research will be using the frameworks of Geert Hofstede on national and organizational culture
(Hofstede, 2014a; 2014b). Previous research indicates that the main reasons for foreign students to
leave Denmark after graduation are the differences in the national -and organizational culture
(Junge & Jacobsen, 2009). The survey questions have therefore been designed in such a manner,
that they touch upon the relevant dimensions in the framework. Geert Hofstede (2014a; 2014b)
argues that each country has different national and organizational cultures, which can resolve in
cultural clashes. Hofstedes framework is therefore relevant to consider when constructing the
research design. Additionally, Maslows theoretical framework regarding hierarchy of needs are the
basis of the dimensions in the conceptual framework and hypotheses (Maslow, 1943).

1.3 Previous research

On the subject of foreign students leaving Denmark, several studies have been conducted prior to
this research. The focus of these studies have been on foreign students in general, whereas this
research focuses on BSS students, hence, the population for this study is further concentrated than
previous studies. However, several important elements from previous and more general studies are
transferrable to this research.
The research suggests that international students create several benefits e.g. socioeconomic gains,
they are valuable to Danish companies and they have a positive effect on the international study
environment on Danish educational institutions (Uddannelses- og Forskningsministeriet, 2013), and
those who leave Denmark after graduation become excellent ambassadors for Denmark (Gaard,
2009).
An analytical report (2009), published by Centre for Economic and Business research education
shows that the number of international students in Denmark has tripled from 1996 to 2006,
supported by the growth in education programs offered in English. The focus of this study was
International students, registered with a CPR-number in Denmark, and thereby excluding those
studying in Denmark for a short period of time. The study shows that international students are
Page 6 of 67

highly overrepresented in specific fields of study, including university educational programs, and
within those especially social sciences. Cand.merc. ranks as the educational program with the
second highest percentage of international students, which is in alignment with the population of
this research (Junge & Jacobsen, 2009). The study shows that the average age of international
students is lower than the average of Danish students and that almost double as many international
graduates in the age group 25-34 are still in Denmark two years after graduation than those in the
age group 18-24. Another factor is civil status, as the study showed that a larger percentage of
international students is single, compared to Danish students and that a larger share of single foreign
students leaves after graduation, than those in a committed relationship (Junge & Jacobsen, 2009, s.
11). Another aspect is current employment, where the statistics on international students from 1999
to 2007, shows that 44 % of those who had a job during their studies, stay in Denmark after
graduation, whereas only 27 % of those without a job stay (Gaard, 2009). Compared to Danish
students, clearly fewer have a job during their studies. The research also suggests that language
skills and a lacking knowledge of the Danish labor market constitutes a barrier in the job search
in Denmark after graduation (Deloitte, 2008).
The migration away from Denmark is dependent on the educational level of the international
students. Research shows that bachelor or master degree graduates within the field of social
sciences are those who are most likely to stay in Denmark, whereas those in educational programs
with a full length of 2-3 years, 3-4, over 5 years and PhD. programs have a higher tendency to
leave. (Junge & Jacobsen, 2009, s. 22). Furthermore, the research suggests that the general wage
level of foreign graduates are higher than the wage level of Danish graduates (Gaard, 2009).
A solution catalogue, published by Deloitte (2008), based on the identification of barriers for
international recruitment suggests that difficulties retaining international graduates in Denmark
stems from them not being able to find relevant jobs. The reasoning behind this is based on a lack of
help during job search, and because of difficulties getting considered for a position, due to the lack
of work experience in Denmark. The identification of barriers in this study also points towards
lower employment rates among international students compared to Danish students during their
time of study (Deloitte, 2008).
The presented prior research in our field of study shows a focus on facts, such as demographic
explanations, as an explanation for the high percentage of international students leaving Denmark
after graduation. Despite the believe that these facts constitute large parts of the explanation, it is

Page 7 of 67

also believed that other factors, such as cultural differences, societal structure, career perspective
and social factors bear a significant part of the explanation, which all are included in the conceptual
framework presented in section 1.6.

1.4 Research Question and overview of research design

Based on the above mentioned background knowledge, this research paper will be conducted with
the following research question in focus;
Which factors influence the motivation of foreign BSS students, to leave or stay in Denmark after
graduation?
The decision process of foreign BSS graduates on staying or leaving Denmark is the central element
in this research. The purpose is to acquire knowledge about which factors that influence the
decision to stay or leave.
For the purpose of a thorough understanding, mixed methods will be used as the overall research
design. The research consists of two studies, one quantitative and one qualitative. The quantitative
study will be the main study and will be executed first. The quantitative study serves to get relevant
information on which factors that influence the expats willingness to stay in Denmark, as a backup
to the main study. The overall research questions will be the same in both the quantitative -and
qualitative research. Thus the method of triangulation will be used.
The following paper is structured as follows: First, hypotheses are formulated. On the basis of these
hypotheses, the conceptual framework is developed. Next, the chosen methods used for both studies
will be elaborated and delineated. Following, there will be an overall discussion and conclusion.
Finally, an alternative use of mixed methods will be discussed.

1.5 Hypotheses
1.5.1 National culture:
The national culture framework has 6 dimensions1. These can be used to analyze the difference
between foreign students native cultures, compared to Danish national culture (Hofstede, 2014b)..
Hofstedes framework (2014b) is relevant for the research question, as it is assumed that cultural

1) Power distance, 2) individualism vs. collectivism, 3) masculinity vs. femininity, 4) uncertainty avoidance, 5) long
term vs. short term orientation and 6) indulgence vs. restraint

Page 8 of 67

differences between Denmark and the respondents home countries, is one of the main factors that
influence foreign students decision to stay or leave after graduation.
Hypothesis 1
The differences in national culture, influences the decision to stay in or leave Denmark after
graduation

1.5.2 Organizational culture:

The organizational culture has 8 dimensions2. These dimensions address inter-organizational
culture, and are used for analyzing the difference between native organizational cultures of both
countries and companies (Hofstede, 2014a). It is expected that differences in organizational culture,
will influence the decision to stay or leave after graduation.
Hypothesis 2
Differences in the organizational culture influence the decision to stay in or leave Denmark after
graduation.

1.5.3 Language barriers:

Language barriers cause misunderstandings, inefficient communication, feelings of alienation,
unintended conflicts etc. In a working environment, these consequences can be problematic for
several reasons. In general, miscommunication is a problem, but especially in a professional
environment, the importance of clear and exact communication is high. Especially within
companies, where the organizational language is also the native language, the risk of foreigners
being left out, feeling unimportant or overlooked is high. Moreover, difficulties communicating in
general, affects the expats ability to create social connections, both inside and outside the
workplace. As social contact is a basic physiological need according to Maslow (1943), language
barriers can have large consequences, for foreigners living in Denmark. Thereby, it is assumed that
language barriers have a crucial role in the decision process.
Hypothesis 3
Language barriers have an influence on the decision to stay or leave Denmark after graduation.

1) Means-oriented vs. goal-oriented, 2) internally driven vs externally driven, 3) easygoing work discipline vs. strict
work discipline, 4) local vs. professional 5) open system vs closed system 6) employee-oriented vs work-oriented, 7)
degree of acceptance of leadership style and 8) degree of identification with your organization

Page 9 of 67

1.5.4 Current employment

It is assumed that having a professional network is an important factor in the job search that is a
natural step after graduation, as it increases the odds of finding a job. Current employment in
Denmark, will consequently to some extend constitute a professional network, that can be beneficial
later on. If the foreign graduate currently has a relevant job in Denmark, it is expected, that it will
contribute positively to the overall desire to stay in Denmark after graduation
Hypothesis 4
Employment situation will influence the motivation to stay or leave after graduation.

1.5.5 Social factors:

As mentioned previously, social connections is a basic human need. The ability to socialize and
create a network is considered crucial to the decision process. An understanding of what defines a
good social network can differ significantly between individuals. Some find family being nearby an
important factor, and others are comfortable with being away, as long as their need for social
connections is satisfied. Thereby this study seeks to get preview of how a good social network, in
the understanding of the individual respondents, affects the decision process.
Hypothesis 5
Social networks will have an effect on the motivation of foreign students to stay in or leave
Denmark after graduation.

1.5.6 The Danish societal structure

Denmark is often referred to as the welfare state, or as part of the Nordic Model, due to the
governments role in protecting and promoting economic and social well-being of its citizens
(Blecher, 2013). This governmental structure can be perceived rather unusual to those who grew up
in a different setting. In the Nordic Model high taxation is redistributed to provide equality of
opportunity and distribution of wealth. Some might have a conservative view on this social model
others might have a more liberal view of the Danish structure (Blecher, 2013). It is expected that
those who are in favor of such model, are more likely to stay in Denmark after graduation, whereas
the likeliness of those who value responsibility for own wellbeing higher than a safety net provided
by the state is assumed to be lower.

Page 10 of 67

Hypothesis 6
The perception of the welfare system will have an effect on the decision to stay in or leave Denmark
after graduation.

1.5.7 Career perspective in Denmark

This parameter reflects the general perception of the respondents assumptions about the work
culture and organizational structure in Denmark, and whether they see themselves as a fit for the
Danish workforce. If Denmark is seen as a potential place to achieve career goals, it indicates that
Denmark in not only found attractive, but also that possible implications, such as language barriers,
to be manageable is considered. Therefore, it is assumed that foreigners, who believe that they will
be able to fulfill their career goals in Denmark, will be more motivated to stay in Denmark after
graduation.
Hypothesis 7
Perceived career potential in Denmark, will have an effect on the decision to stay in or leave
Denmark after graduation.

1.6 Conceptual framework

Based on the seven hypotheses mentioned above, a conceptual framework has been drafted (Figure
1). The purpose is to illustrate the researchers prior understanding and perception of the
phenomenon of foreign graduates decision to stay in or leave Denmark after graduation. Before the
phenomenon is explored, seven factors are expected to have a significant influence. The direction
and importance of the effect from these factors are unknown, but will either reduce or increase the
motivation for staying in Denmark.
The results from the data analysis will contribute to a refined and improved conceptual framework,
that will presented in section 4.2. The hypotheses stated above will be tested, and the verification or
falsification will be the base, wherefrom the factors will be corrected.

Page 11 of 67

Figure 1

1.6.1 Expats, studying on AU, School of Business and Social Sciences:

As mentioned (Section 1.1.2), an expat is defined as; a student from a foreign country that has
chosen to live and study in a country different from their home country, for the purpose of this
study. Different perspectives associated with different fields of study, and the future possibilities for
working in Denmark vary among the different fields. The population in this research is limited to
foreign students at Aarhus University, School of Business and Social Sciences, as this faculty has
the highest number of foreign students, and business students are found to be most relevant study
area, for this research.

1.6.2 Factors that influence the decision to stay or leave:

It is believed that the above mentioned factors have a significant influence on the decision process
for expats, to stay in or leave Denmark after graduation. A positive effect is defined as a factor
increasing the desire to stay after graduation. A negative effect is defined as a factor that decreases
the desire to stay after graduation.

Page 12 of 67

1.6.3 The plus/minus arrow:

The purpose of the arrow is to indicate how the different factors affect the expats decision to stay
or leave. The effect will have a positive, neutral (insignificant) or negative effect on the decision
process. It is expected that the effect of each factor can be determined from the collected data.

1.6.4 To stay or leave scale:

This scale shows the likelihood of a specific student staying in or leaving Denmark after graduation,
taking all significant factors into consideration, and is thereby determined by the multiple regression
model from section 2.5 supplemented by results from section 3.2. The results will show which
factors are significant to the overall model. The conceptual framework shall be seen as a starting
point for the final multiple regression model that will be the result of the study. It is noted that
several of the above mentioned parameters, might not have a significant influence on the foreign
students decision to stay in or leave Denmark after graduation, and will be excluded from the result
if proven insignificant.

2 Primary quantitative study

2.1 Quantitative study - Method
In the following section, reflections and decisions regarding the quantitative study will be
presented. The study will be conducted using Google Consumer Surveys3, which is a free online
tool, most convenient and suitable for this study. This program is manageable, easy to use, and
provides the possibility of transferring data output into a variety of statistical programs. The
quantitative research will have a sample size of 62 respondents.

2.1.1 Research guide structure:

The research questions are derived from the conceptual framework. The questionnaire used can be
found in Appendix 22. The questions are designed in a manner that is believed to provide an output
that will prove or reject the hypotheses. The questionnaire is divided into sections based on the
conceptual framework, and is primarily consisting of close ended questions, except three questions
with the purpose of categorizing the respondents into different groups4. Close ended questions are

http://www.google.com/insights/consumersurveys/home

How old are you? What is your home country? What program are you studying?

Page 13 of 67

used to allow comparisons and generalizations of the answers in the statistical analysis. In order to
identify possible respondents for the secondary study, the possibility of entering contact information
is provided in the end of the primary study. It is assumed that those who provide their contact
information has an interest in the field of study, and find themselves able to provide relevant
insights on the subject. The questions will primarily be Likert scaled questions, with options
between one and seven. Binary questions, such as gender, are incorporated, primarily in the general
section of the questionnaire.

2.1.2 Statistical analysis:

The quantitative data collection will generate a large number of variables. In order to reduce this
number, and making the data suitable for further analysis, a factor analysis is conducted. The
purpose is to transform variables into clusters, in alignment with the conceptual framework. The
factors will be the independent variables within the following statistical analysis; a multiple
regression. A multiple regression model can test the seven different hypotheses, and check whether
or not they have a significant influence on expatriates motivation for staying in Denmark. These
statistical analyses will be presented in the following sections.

2.1.3 Choice of scale intervals:

All interval scaled questions ranges from one to seven. This range is constant throughout the
questionnaire to avoid unnecessary confusion. The meaning of each point on the scale differs,
depending on the specific question, and is elaborated in each question. A seven point scale provides
the possibility to express ones opinion in greater detail compared to a five point scale, while still
offering a neutral possibility. Thereby respondents will not feel obligated to provide an answer on a
question that they are neutral towards.

2.1.4 Sampling technique:

The random sampling technique is the standard sampling technique used in quantitative studies
(Miles et al., 2014). This technique strengthens the validity of data outputs, and ensures an unbiased
selection of respondents.

Page 14 of 67

2.1.5 Distribution of the questionnaire:

The questionnaire will be posted on various groups and communities on social networks5, for
foreigners in Denmark. The target group will be explained, and several questions in the
questionnaire are designed to ensure respondents within this target group. Those answers that fall
outside the target group are excluded from the research.

2.1.6 Pilot testing:

Pilot testing is used to strengthen the validity of the quantitative research. By conducting a pilot
test, it is possible to intercept errors and address points of confusion for the test group, before the
actual study is conducted. The respondents of the pilot test are a combination of Danish and
International students, found to have the relevant knowledge to provide feedback on mistakes,
unclear questions and design errors in the questionnaire.
By using pilot test respondents, who can be categorized as similar to the target group or the
researchers, the overall validity of the research will be strengthened (Fink, 2006).

2.2 Factor analysis as a method:

The output from the primary study shows 34 different variables from 62 respondents. Most
variables are Likert scaled, and some are based on a nominal level of measurement. In the collected
data, subsets of variables with significant correlation are expected. On the other hand, significant
heterogeneity among the subsets are also expected. When clustering factors, Tabachnick and Fidell
(2012) mention two major methods; Factor Analysis (FA) and Principal Component Analysis
(PCA). The distinction between FA and PCA are that FA is used when the intention is to get a
theoretical solution, and PCA is used when the intention is to get an empirical summary of the data.
The purpose of this analysis is to develop a model that can explain some of the variation in the
motivation for staying or leaving Denmark after graduation, thus getting a theoretical solution,
which makes FA the right tool for this. Factor analysis is also chosen to reduce the number of
variables and derive factors for later use, and are used to test the conceptual framework, identifying
influential factors.
In this stage of the research project, an exploratory factor analysis will be performed. The
underlying processes are relatively unknown, and the factor analysis is expected to enhance the
understanding of the phenomenon.
5

Facebook and LinkedIn

Page 15 of 67

2.2.1 Exclusion of variables based upon unfulfilled assumptions

Before a detailed review of the factor analysis, the process of excluding variables is examined. This
exclusion is based on the rules of thumb, used in factor analysis. This progress is chosen to align
with the process of a factor analysis.
The anti-image correlation matrix (Appendix 5) shows that four variables6 have relatively low MSA
values in the diagonal, where the variable Fair_tax scores lowest with a value of 0.319. The rotated
component matrix (Appendix 6) reveals that Fair_tax does not fit into any factor. This variable has
a low loading on Factor 2 and 4, while it has a high negative effect on Factor 6, which is not
pragmatic. Fair_tax is thereby considered an outlier and is excluded from the analysis. Next, an
issue of cross loadings on two variables arises7 (Appendix 7). Both variables have a relatively high
cross-loading, and load high on Factor 1 and 6. Pragmatically both variables fit together in Factor 1,
and the variables are kept in the analysis for now. The Anti-Image matrix (Appendix 8) reveals that
dk_individual has a low MSA value of 0.343 and therefore this variable is excluded instead. Now
only one variable8 is cross loading (Appendix 9), thus the next variable excluded from the analysis.
The variables are now all clustered in six factors, with significant correlations and an absence of
cross loadings (Appendix 10).

2.2.2 Number of factors

Three methods are appropriate for deciding the amount of factors included in an analysis,
Eigenvalues, Scree Plot and the cumulative percentage in the Rotation Sums of Squared Loadings
matrix.
According to total of initial Eigenvalues into account, it can be concluded from the Total Variance
Expected (Appendix 18) table, that the final amount of factors derived from this factor analysis
should be six, because this is the amount of components with a total eigenvalue above one.
Another method to determine the amount of factors that should be included in the analysis is
looking at the Scree Plot (Appendix 19). At the point where the curve flattens, the amount of factors
on the left side should be the amount included in the factor analysis, which in this case are five.
The third option is determining the amount of included factors based on the cumulative percentage
of the Rotation sums of Squared Loadings (Appendix 18). Using this method, the optimal amount of
variables are determined by the amount of factors equal to the point where the cumulative rotation
6

Dk_individual, lang_barrier_job, Residence_prob and Fair_tax

Career_dk and Career_pot_dk
8
career_dk
7

Page 16 of 67

sum of squared loadings exceeds 60 %. Also this method indicates that five factors should be
included in the analysis.
Two out of three methods indicate that five factors is the optimal solution, but after evaluating the
model with five factors, two variables did not load on any factor. To solve this problem, the variable
Gov_benefit was excluded. After this step, the scree plot and the cumulative percentage of the
Rotation sums of Squared Loadings showed that the number of factors should be reduced to four.
By reducing the number of factors to four, even more variables start to load less on the specific
factors, and the result lacked of logical composition of the factors. The combination of six factors
are found to be most appropriate, as the logic fails when factors are removed, and is thereby chosen
as the number of factors for further analysis.

2.2.3 Rotation
After the extraction of the variables, a rotation is used to maximize high correlation between
variables and factors, and to minimize low correlations. There are several different rotation
methods, categorized as either orthogonal or oblique. Here, the decision is between Varimax
(orthogonal) and Promax (oblique). The Varimax maximizes the squared loading variance across
variables. Promax then simplifies the results of the Varimax rotation by making small loadings
closer to zero, and maximize high loadings. This process starts with the Varimax rotation, and then
each axis is rotated one by one.
During the factor analysis, both rotation methods are tried, to identify possible differences.
According to the correlation matrix (Appendix 3), there are a relatively high number of correlations
above 0.32, which is the threshold for Tabachnick and Fidell (2012) because the overlap in the
variance will exceed 10 %. Therefore the Promax rotation is used. When consulting the component
matrix (orthogonal) and the pattern matrix (oblique), there are no clear differences. Some loadings
increased a little, and some decreased. This suggests no reason for using an oblique rotation,
because it does not simplify the results, and the Varimax rotation is thereby used in this factor
analysis.

2.2.4 Factorability of R (after variable/factor reductions)

According to Tabachnick and Fidell (2012), the Kaisers measure of sample adequacy has to exceed
0.6. However, 0.5 is a commonly used limit. After removing insignificant variables, the total
intercorrelation between the variables are 0.592, which is considered acceptably close to 0.6.

Page 17 of 67

The Bartels significance test of zero correlation between variables are falsified, with a level of
significance on 0.000. To support this, the approximate chi-square (304.3) is almost three times
larger, than the degrees of freedom (105). This aligns with the conclusion of Kaisers Measure of
Adequacy, where the requirement is almost fulfilled.
The correlation matrix (Appendix 3) reveals several correlations among items. An acceptable
number of correlations exceed 0.3, and therefore a pattern is anticipated.
The values in the off-diagonal in the anti-image correlation matrix (Appendix 8) contain small
correlations, which is a requirement for a good factor analysis.
The overall assessment of the assumption regarding factorability of variables are seen as satisfied.
Even though the two tests are a little below the limit, there are also strong signs of correlations,
which makes a factor analysis possible.

2.2.5 Assumptions for the best fitting model of factors:

The following assumptions are based upon the best fitting model (BFM), with 6 factors and without
the variables; fair tax, dk_individual and career_dk.
Sample size and inadequate data:
It is recognized that the sample size used in the following project does not meet the requirements set
by MacCallum et al (1999), who state that the optimal sample size for a factor analysis should be
100-200 respondents.
However, the sample size is not considered a problem, as the population is 571, our sample size of
62, approximately 11% of the population, will be representative and able to explain the
phenomenon. Since the sample size is below 100 it is recommend that all correlations are greater
than 0.6.
All communalities, except Job_odds_DK and career_pot_dk (Appendix 3), are higher than 0.6.
Thus this assumption is not fully met, which will decrease the validity of the data. However, all the
variables in this step are kept for further analysis.
Multivariate Normality:
It is assumed that all the variables are normally distributed. This is addressed by the skewness and
kurtosis (Tabachnick & Fidell, 2012). It is impractical to conduct a normality analysis on all
variables, thus only variables9 suspected not to follow this distribution are chosen for further
investigation. The test of normality (Appendix 1) shows that the chosen variables are influenced by
9

Dif_org_Structure, Flat_structure and Soc_network_dk

Page 18 of 67

left and right skewness, thus not normally distributed. However, it is assumed that the remaining
variables are normally distributed, and skewness on some variables was expected. Failing the
normality tests reduces the validity of the analysis (Tabachnick & Fidell, 2012). However, the
variables will be kept for further analysis.
Multivariate Linearity:
A linear relationship between pairs of variables in the data set is assumed. Due to the amount of
variables it is insufficient to check all variables. Therefore spot checks are made on those variables
that could create problems.
The test of linearity between variables (Appendix 2) shows that the multivariate linearity between
the variables Dif_org_Structure and Flat_structure is not optimal, but perceived acceptable. The
test (Appendix 2) furthermore shows a multivariate linearity between the variables Danish_skills
and pro_danish. Thereby linearity issues between some of the variables are identified, thus this will
have a negative effect on the overall validity. However there is no clear indication of curvilinearity,
which degrades the analysis and the variables, will be kept for further analysis (Tabachnick &
Fidell, 2012).
Absence of Outliers among cases:
The conducted data does not reveal any significant outliers. In a Likert scaled survey it is
expectable that some respondents choose the easy answer, by checking the same answer to all the
questions. However, this has not been a problem, and there have been no contradicting answers.
Thereby no outliers were found in the data, and this assumption is fulfilled. The absence of outliers
can be explained by the respondents perception of the topic as relevant to them.
Absence of multicollinearity and singularity (after variable/factor reductions):
In order to check if some of the remaining variables should be removed, a test of multicollinearity
and singularity is made. This will be tested by examination of the SMCs of all variables where they
serve as DV to all other variables as IVs (Tabachnick & Fidell, 2012). In the communalities matrix
(Appendix 20) the SMCs of the variables are shown. There are no signs of singularity, where a
variable is equal to one. There is no clear indication of multicollinearity either. According to
Tabachnick and Fidell (2012) the SMC value needs to be dangerously close to one before there is
multicollinearity. In this case the highest value is 0.848, which is not considered dangerously
close to one. Therefore, the assumption of absence of multicollinearity and singularity are fulfilled.

Page 19 of 67

Absence of outliers among variables

To check for outliers among variables the SMC values (Appendix 20) are used (Tabachnick &
Fidell, 2012). There is no sign of outliers among variables, since the lowest SMC value is 0.549.
This is a clear indication that outliers among variables are not an issue in this study.

2.3 Results of factor analysis

In the previous sections, it has been determined, that the optimal number of factors are six. Below is
an overview of the factors and the variables within:
Table 1

The names of the factors are based on pragmatism, or relating to the variable with the highest
loading within the factor, in accordance with the recommendations of Tabachnick and Fidell
(2012). In table 1 the variables are ordered by loadings, which provide an overview of which
variables having the highest influence on each factor.
The final factors are first of all: the product of significant correlations between the given variables
forming each factor. However the pragmatic logic behind factoring the variables are equally
important (Tabachnick & Fidell, 2012). Table 1 show that the variables within each factor are in
alignment with these prerequisites. Fac_1 include variables describing the Danish organizational
structure. Factor two, Fac_2, is based on variables defining the respondents network, in correlation
with the chances of finding a job. Factor three, four and five consists of only two variables each.
Aside from that particular issue, Fac_3 is based on the respondents Danish language proficiency,

Page 20 of 67

in relation work environments with Danish as the main language and in general. Fac_4 is based on
barriers met when applying for a job in Denmark. Fac_5 is more complex than the others, since the
factor describes benefits from a job in Denmark and career potential, hence, the name Job related
attributes in Denmark. Fac_6 includes complications of getting residence permits, based on
experience and expectations.
Table 1 reveals that three variables; fairness of the tax in Denmark, Danes are closed and
individualistic and how likely do you think it is that you can fulfill your career goal in Denmark,
do not have a significant correlation to any other variable, and are thereby not part of a factor. From
a pragmatic viewpoint, it can be argued that these variables should align with one of the above
mentioned factors, but the test shows no significant correlation, and they are thereby excluded from
further analysis. The identified factors are named in accordance with Table 1, and will be used for a
multiple regression.

2.3.1 Discussion of the factors and their connection to the conceptual framework
In the previous sections a six factor solution is argued to be the most suitable. This section will take
a more critical approach to the solution and the construction of factors.
Three out of five factors are based on two variables, which can question the reliability of these
factors. According to Fidell and Thabinick it can be hazardous to use factors based on two
variables, but this claim depends on the pattern of correlation between the variables. The correlation
matrix shows a strong correlation between Pro_danish and danish_skills, which in addition
correlate at a minimum with the other variables. Factor 4, 5 and 6 are less reliable, because the
correlation is relatively low between the clustered variables. These factors, and especially Factor 5,
represent some interesting and unexpected findings, and provides valuable knowledge. Therefore,
the factors are kept in the analysis, despite that the reliability is uncertain.
The discussed assumptions in section 2.2.5 have shown minor problems regarding multivariate
linearity, multivariate normality and the sample size ratio. As a consequence of these issues, the
overall validity is decreasing. This means that the correlation between the hypotheses lack a
compelling significant support. A larger sample size could however eliminate the minor assumption
issues in the factor analysis.
When assessing the loadings, the standard rule of interpreting variables, indicates that loadings
exceeding 0.32 should be interpreted. This indicates more than 10 % overlap in variances between
the concerned variable and the rest. This factor analysis contains only four loadings that breach this
rule of thumb. This is satisfactory, and indicates a large amount of heterogeneity between factors.
Page 21 of 67

In this factor analysis 0.45 is chosen as the lowest interpreted loading (Appendix 7). The reason for
this pragmatic choice is the relatively high loading on other variables, and a good reasoning behind
the division of variables. Except these two loadings career_pot_dk and career_dk, the lowest
loading is 0.595, which is considered good, with an overlapping variance above 30 %. A large part
of the loadings exceed .70, which is considered excellent (Tabachnick & Fidell, 2012). Therefore
the homogeneity in each factor is assessed to be acceptably high.
In the following section it will be argued how the factors, found in the factor analysis, can be
connected to the hypotheses presented in section 1.5.
Table 2

Factor 1 - Organizational structure

The variables in this factor emphasize an appropriate label as organizational structure. The
variables in factor one aligns with Hypothesis 2. It is recognized that power distance normally is
associated with national culture. However, in this context it is argued that power distance should be
understood in connection with a flat hierarchical organizational structure10. Thus this factor will be
used to confirm or reject Hypothesis 2, in the multiple regression analysis.
Factor 2 - Network
This factor is best described with the label network, simply because the two variables with the
highest loadings are concerning networks. It is argued that this factor is connected to hypothesis 4,
10

Question connected to Power_dist: There is a relatively low power distance in Danish companies: (Power distance
refers to how formal and big the distance is form the manager to the employees

Page 22 of 67

current employment, hypothesis 5, social factors and hypothesis 7, career perspective. In

accordance with the assumptions made in the introductory section, these variables are closely
connected to each other and influence each other in the same direction.
Factor two will thereby be used to conclude, whether there is a support for hypothesis 4, 5 and 7, in
the multiple regression analysis.
Factor 3 - Language skills
This factor is labeled in connection to hypothesis 3. The variables in Factor 3 clearly align with
hypothesis 3, concerning how language barriers might influence the research question. Thus, Factor
3 will be used to conclude whether there is support for hypothesis 3, in the multiple regression
analysis. As already mentioned, it is recognized that the factor consists of only two variables and
might influence the significance of the factor.
Factor 4 - Organizational cultural barriers
The variables in factor four can be categorized as cultural organizational barriers. These variables
explain how cultural differences can create barriers when entering the Danish workforce. Even
though it is the context of an organization, the true nature of the problem is argued to be connected
to national culture. Thereby this factor can be connected with hypothesis 1. Factor four also consists
of two variables, which is inadequate and can influence the significance of this factor in the
following analysis. Factor four will be used to confirm or reject hypothesis 1, in the multiple
regression analysis.
Factor 5 - Job related attributes in Denmark
No hypothesis stated in section 1.5 is in alignment with this factor. However, the factor is still
analyzed through the multiple regressions. If the analysis concludes that this factor is significant,
the conceptual framework has to be modifies.
Factor 6 - Residence problems
No hypotheses have a connection to this factor either. The approach to this factor is identical to
factor five. Residence problems are a fitting label for Factor 6.
As no hypothesis can be connected to factor five and six, it indicates faults in the conceptual
framework. Possible modifications to the conceptual will be made, if these factors show a
significant influence on the dependent variable. If factor 5 or 6 are showed to be insignificant, no
changes will be made in the conceptual framework.

Page 23 of 67

2.4 Multivariate Regression analysis method.

The hypotheses of this research are that certain factors influence expatriates decision of leaving or
staying in Denmark, after their graduation. In order to test if there is a correlation between the six
factors, found in the previous analysis, and the dependent variable Motivation to stay in Denmark, a
multiple regression analysis will be conducted. A multiple regression model seeks to find
correlations between dependent and independent variables, in order to test how the independent
variables influence the dependent variable.

2.4.1 Assumptions
In order to assure validity in the results it is crucial that certain significant assumptions are fulfilled.
Tabachnick and Fidell (2012) mention seven different assumptions required for a multiple
regression analysis. In the following section these seven assumptions will be elaborated.

2.4.2 Ratio of Cases to Independent variables

Tabachnick and Fidell (2012) recommends that the number of cases need to exceed 50 + 8m (m=#
of IVs), when testing for multiple correlation, and 104 + m, when testing individual predictors
(Tabachnick & Fidell, 2012). As the purpose is to test correlation between the six factors, from the
factor analysis, and the dependent variable, the number of cases in this research should exceed
50 + 8 6 = 98 in order to test for the overall multiple correlation and 104 + 6 = 110 in order to
test on the individual predictors. The sample size used in this research is 62, which means that this
requirement is not met. Having many independent variables compared to the sampling size can
result in unrealistic predictions of the correlation between the IVs and the DV (Tabachnick &
Fidell, 2012). As argued above the sample size is assumed sufficient, but this assumption breach
should be considered when the results are interpreted.

2.4.3 Absence of Outliers among the IVs and on the DV

It is crucial to spot outliers in the data set, since extreme cases impact the regression model, and
thereby mislead the results. Extreme cases will influence directly on the coefficients and create bias
(Tabachnick & Fidell, 2012). Biased coefficients are not sufficient when generalizing the
population.
Its possible to test for outliers either pre or post the regression run. Tabachnick and Fidell (2012)
recommends screening of outliers prior to the regression run, while testing for outliers after the
regression run, might cause over-fitting.

Page 24 of 67

The histogram of independent variables (Appendix 12) shows indications for minor outliers within
Factor 1, 5 and 6. In all three factors, few cases are not connected to the other cases in the
histogram, and are located far from the mean. This can be a symptom of an inadequate sample size,
or a sign of outliers within the data. Outliers found within the independent variables are considered
insignificant and will therefore not be excluded from the analysis. Again, this breach in assumptions
will be considered in the results. The histogram for the dependent variable (Appendix 17) shows a
normal distribution, with no signs of outliers. Thereby the significant assumption of absence of
outliers among the dependent variable is satisfied.

2.4.4 Absence of multicollinearity and singularity

In order to avoid bias, it is vital to limit the presence of multicollinearity. In this study the
independent variables are generated in the factor analysis, where singularity and multicollinearity
are taken into account. The factor analysis indicated strong correlations between several variables.
These variables are transferred into factors, to become suitable for this regression analysis. By
clustering strong correlated variables, and secure a minimum of cross loadings between the
variables, multicollinearity and singularity is reduced. Thereby this assumption is assumed to be
met.

2.4.5 Normality, Linearity, and Homoscedasticity and independence of Residuals

i ~ NF : An assumption for a multiple regression is a normal distribution, as deviations can

indicate outliers in the data (Tabachnick & Fidell, 2012). The histogram of residuals (Appendix 14)
shows relative normality where the variation can be explained by the sample size.
Furthermore, the single residuals of the dependent variable in a scatter plot (Appendix 15), shows
that the normality of the residuals are acceptable. There is a fine concentration of residuals close to
0, and the frequency of residuals decrease moving away from zero.
When checking for linearity in the scatter plot of residuals (Appendix 16), the residuals do not show
any clear signs of nonlinearity. Furthermore, the linearity of residuals plot (Appendix 16) shows a
relatively fair linearity, with only smaller indications of outliers, but none extremes. Thus, the
assumption of a linear relationship between the residuals and the predicted value is perceived
satisfactory.

Page 25 of 67

V ( i ) 2 : The scatter plot of residuals (Appendix 15) reveals that the variance of the residuals are

constant, thus homoscedasticity. A clear sign of heteroscedasticity is when moving along the x-axis;
the band of the residuals becomes wider. This is not a problem in this data set, thus the assumption
is satisfied.
Corr ( i , j )) 0 : The scatter plot of the residuals (Appendix 15) does not indicate any clear sign of

correlation between the residuals. This means that the residuals are independent of each other. The
assumption of independent residuals is thereby fulfilled.
The scatter plots do not indicate any clear assumption breaches regarding residuals. However, the
small sample size is compromising the overall compliance. A larger sample size is expected to have
provided more clarifying results, and will be taken into consideration in the results. Overall, the
assumptions of normality, linearity, homoscedasticity and independence of the residuals are
assumed to be satisfactory.

2.4.6 Absence of Outliers in the Solution

The scatterplot of the residuals (Appendix 15) does not indicate any outliers. A known explanation
for this could be the small sample size, which reduces the probability of outliers. Outliers will have
higher impact of the results on smaller sample sizes. When the sampling size increases, the risks of
outliers will raise, while the impact will decrease. However, in this research outliers are not an
issue, and are not a concern when analyzing the results.

2.4.7 Excluding variables using the enter method

When interpreting the six factors into a regression model, with the dependent variable Motivation to
stay in Denmark, only few factors are found to be significant. There are several methods to exclude
the insignificant variables within the regression analysis, whereas the most common and accurate
option is the Enter method (Tabachnick & Fidell, 2012). This method is quite forward - first all
variables are interpreted and insignificant variables are manually extracted one by one, until all
variables within the model are significant. The coefficients matrix (Appendix 13) shows the output
of all six factors. The matrix (Appendix 13) shows that the least significant factor is Factor 6, with a
p-value of 0.738, thus the first factor removed from the model. Thereafter the process is repeated.
The final regression model (Appendix 11) reveals that only two factors, Factor 1 and Factor 5, are
significant, with p-values of respectively 0.00 and 0.017. Factor 1 and 5 are thereby the two

Page 26 of 67

significant independent variables that can explain the expatriates motivation to stay in Denmark, at
least to a certain degree.

2.5 Results from final multiple regression model

The multiple regression model for this study is as follows:
_ = 4.629 + 0.843 _ + 0.431 ____ +

The results of the regression model show that both factors have a positive influence on the
expatriates motivation to stay in Denmark. This means that both the Danish organizational
structure and the job related attributes motivates expatriates to stay in Denmark.
The motivation is in average 4.629, which is 0.629 above the neutral answer of 4; this indicates that
the expatriates in Denmark initially are motivated to stay in some degree. Seen from the
coefficients, it can be concluded that the influence of the Danish organizational structure is most
important for the expatriates motivation to stay in Denmark.

2.5.1 Adjusted R Square

Table 3 shows the overall fit of the regression model. R square shows to what degree the two
factors explain the dependent variable. However in a relatively small sample size, the Adjusted R
Square measure is a more precise estimate for the population (Tabachnick & Fidell, 2012). Smaller
sample sizes tend to overestimate the fit of the model (R square), thus the adjustment of R square.
The adjusted R square for this multiple regression model is 0.310, which means that Fac_1
Organizational_structure and Fac_5, Job_related_attributes explain 31 % of the expatriates
motivation for staying in Denmark.

Page 27 of 67

Table 3

2.6

Discussion/perspective

The assumptions discussed for the regression model (Section 2.4.1) had some issues, which
influence the results. The breach on the sample size assumption compromises the validity of the
coefficients, and they can thereby be biased. Therefore it is not significantly safe to conclude on the
independent coefficient values within the regression model. However, what this analysis with
certainty can reveal is that Factor 1 and 5 has a positive impact on the dependent variable.
An interesting result of the regression analysis is that there are no significant factors that have a
negative impact on the motivation for staying in Denmark. This shows that some aspects might not
have been considered in the data collection process. Since it pragmatically makes no sense, that
there is an issue of high percentage of expatriates leaving Denmark, while there is no significant
factors pushing them away.

2.7 Results
Hypothesis 1
The differences in national culture, influences the decision to stay in or leave Denmark after
graduation
This research does not find significant support for Factor 4, and hence no support for Hypothesis 1.
Therefore, organizational cultural barriers cannot be concluded to have a significant effect on the
motivation to stay in Denmark. As mentioned above, Factor 4 consists of only two variables
(Appendix 10), which can be an explanation for the insignificance of this factor.
Hypothesis 2
Differences in the organizational culture influences the decision to stay in or leave Denmark after
graduation

Page 28 of 67

The multiple regression model finds Factor 1, Organisational_structure, significant. Therefore, it

can be concluded that the Danish organizational structure has a significant influence on foreign
graduates motivation to stay or leave after graduation. The regression coefficients of Factor 1 show
a positive effect, and Hypothesis 2 is confirmed. The factor coefficients emphasizes that the flat
organizational structure, currently categorizing many Danish firms, motivates foreign graduates to
stay after graduation. In other words, research suggests that Danish firms would not gain any
benefits by changing their organizational structure, in the pursuit of foreign graduates.
The variable Wage_DK is a part of Factor 1, can offer part of the explanation for factor ones
positive influence on the desire to stay, as the general wage in Denmark is higher than most other
countries and as the wage level of foreign graduates from non-EU countries 5-8 percent higher than
the Danish graduates (Junge & Jacobsen, 2009; Hansen, 2010).
Hypothesis 3
Language barriers have an influence on the decision to stay or leave Denmark after graduation.
The multiple regression analysis shows no significant support for Hypothesis 3. The P-value of
Factor 3 is above the 0.05 limit, thus this research concluded that language barriers do not have any
significant influence on the decision to stay or leave after graduation. This research therefore
indicates that language skills of foreign students do not significantly influence their decision to stay
or leave. Similar to Hypothesis 1 an explanation could be that Factor 3 only consists of 2 variables
(Appendix 10). Another possible explanation is that the theoretical background used for
constructing Hypothesis 3 is weak
Hypothesis 4, 5 and 7
Employment situation will influence the motivation to stay or leave after graduation.
Social networks will have an effect on the motivation of foreign students to stay in or leave
Denmark after graduation.
Perceived career potential in Denmark, will have an effect on the decision to stay in or leave
Denmark after graduation.
As stated in above, Hypothesis 4, 5 and 7 are related to Factor 2 Network. The multiple regression
model concludes, that there is no significant support for Factor 2, thus no support for Hypothesis 4,
5 and 7. This research therefore implies that current employment and the networks associated with

Page 29 of 67

that do not have a significant influence on the decision to stay or leave. Furthermore, the
insignificant results of Factor 2 suggest that the belief of reaching career goals has no significant
influence on the decision to stay or leave after graduation.
Hypothesis 6
The perception of the welfare system will have an effect on the decision to stay in or leave Denmark
after graduation.
As mentioned before, is the conceptual framework is not associated with Hypothesis 6 through any
factor found identified in the factor analysis. The multiple regression model does not imply any
significant influence of a positive attitude towards the welfare state on the motivation of expats to
stay in Denmark.
Factor 5
As previously mentioned, there is no hypothesis connected to Factor 5. However, the multiple
regression analysis concludes that Factor 5 is significant. This research therefore concluded, that
Job related attributes in Denmark regarding career potential, paid sick leave and vacation have a
significant influence on the decision to stay or leave. Since no hypothesis can be fully argued to
have a connection to Factor 5, it is concluded that this is a missing aspect in the conceptual model.
Factor 6
No immediate connection between Factor 6 and the drafted hypotheses is found. As this factor was
found insignificant in the multiple regression model, this factor will not lead to changes in the
conceptual framework. Thus this research conclude that problem related to previous as well as
expected future residence problems, does not have any significant influence on foreign graduates
decision to stay or leave after graduation.

2.8 Conclusion and modification to the conceptual model

The research concluded that the factors, which influence a foreign graduates decision to stay or
leave after graduation, are organizational structure and job related attributes in Denmark. The
analysis further conclude that with the multiple regression model (section 2.2), this study is able to
explain 31% of the phenomenon. The research further concluded that there are no significant
connections, between Hypothesis 1,3,5,6 and 7, and the factors that influence the decision to stay or
leave graduation.

Page 30 of 67

This conclusion consequently implies that the current conceptual model (figure 1), has to be
modified. Firstly, all factors that have been proven insignificant have to be removed, since this
research concludes that these do not provide significant answers to the research question. Secondly,
since Factor 5 is found significant, a new dimension, Job related attributes in Denmark, arises in
the conceptual framework. Both dimensions have a positive effect on the motivation to stay in
Denmark, which will be illustrated in the final model.

3 Study 2 Qualitative
In order to provide the most suitable understanding of the research problem, the following analysis
is based on a mixed method study, which favors using quantitative and qualitative data. A
combination of these approaches, allows for different assumptions, methods of data collection and
analyses. A collection of both: closed-ended quantitative data and open-ended qualitative data prove
advantageous to better understand a research problem (Creswell, 2014).
The primary study revealed the following significant factors: FAC1_Org_struct and
FAC5_Job_related_atrb_DK. These factors can explain 31 % of the expatriates motivation to stay
in or leave Denmark after graduation. The purpose of the secondary study is to gain further
collaboration on the issues found in the quantitative study, thus using a triangulation approach.

3.1 Method
3.1.1 Interview structure:
The form will be semi-structured and the interview will be conducted with one interviewer and one
interviewee. The semi-structured approach then provides the opportunity to further investigate the
relevant topics and issues found in the quantitative study. In order to store data the interview will be
taped and transcribed. The expected duration of the interview is approximately 25-40 minutes.

3.1.2 Question development:

The secondary study is conducted as a qualitative study, and is highly dependent on the result of the
quantitative research. The personal interview will be conducted with a randomly selected
respondent from study one (Miles et al., 2014). Miles & Huberman (2014) suggest that the optimal
approach to an interview is to begin with general practical questions. Subsequently questions related
to the research topic will be asked, divided into theme based sections (Appendix 21). The research

Page 31 of 67

questions were developed with the purpose of providing an insight into the conceptual framework
and to verify or dismiss the hypotheses.
This data collection method is chosen, as it is believed that this method will produce the most valid
data output. The personal interview is chosen in order to get a deeper insight into the results from
the primary quantitative study. Personal interviews provide the possibility for detailed answers, and
thereby gain information to support significant results and provide answers for insignificant results
from study one.
The provision of a variety of investigation approaches within the same phenomena, allows the
authors to adapt a triangulated approach, which increases the validity of the study. In this
research, triangulation is achieved through the use of a survey, followed by a qualitative interview.
The analysis will then evade limitations associated with a single method. Moreover, triangulation
offers the prospect of enhanced confidence in findings by repeated verification provided by
different methods of measurements (Miles et al., 2014).

3.1.3 Interview form and choice of interviewee:

To ensure the comfort of the respondent, the researcher with a native background most similar to
the respondent is chosen as the interviewer. This interviewer is chosen to create common ground,
ease tension and create the possibility of complete honesty, without the fear of insulting the
interviewer, as the Danish nationality and culture is part of the field of study. It is important that the
interviewer manages to come across as trustworthy and well informed, without intimidating the
respondent. Furthermore the interviewer has to act neutral, while showing interest in the
respondents answers, so motivate the respondent without influencing the answers, in order to
enhance the validity.
By easing this tension, the validity of data will improve. To ensure the comfort of the respondent,
the purpose of the interview is elaborated. Furthermore, the respondent is introduced to the
interview guide, and asked permission for recording the interview. Lastly the respondent has the
opportunity to ask general questions. It is recognized that there are drawbacks associated with this
method of research, such as the interviewee affecting the respondents, resulting in a bias data
output, but as described above, the interviewer is chosen to minimize these issues. The interviewer
will ask for clarification or examples in situations where the respondent is addressing relevant
issues. Furthermore, the interviewee will strive to read the body language of the respondent, in
order to recognize misinterpretations and confusion.

Page 32 of 67

3.1.4 Analysis of data:

The personal interviews will be recorded, with permission from the respondent. The interview will
be transcribed and subsequently coded in order to find patterns. The hypothesis coding method is
used for the qualitative analysis (Miles et al., 2014). This method aligns with the overall
triangulations approach for this research. This method is chosen because it enables this research to
investigate the same phenomenon with different methods. Thereby the codes are constructed before
the actual interview is executed. In order to corroborate the presented hypotheses, the pre-structured
codes are named after the seven hypotheses. This enables the possibility to either confirm or reject
the hypotheses. This further enables a possibility to compare what hypotheses each study find
significant. When the transcription is coded each appearance of the different codes is counted.
Furthermore, it is examined whether any answers are influenced by more than one code. This is
presented in an appearance and correlation matrix.
The triangulation approach is then used to compare whether the two studies find the same
hypotheses significant. If not, the overall validity of the researchs conclusion will decrease.
It is recognized, that the multiple regression found that Factor 5 is significant, but as it is not
connected to any hypothesis and the triangulation method is used, issues connected solely to Factor
5 will not be further investigated.

Page 33 of 67

3.2 Results
Table 4

Table 5

(See Appendix 21 for transcript)

Hypothesis 1
As mentioned in section 2.7, the quantitative study does not find significant statistical support for
Hypothesis 1, as there is no significant support for Factor 4. The respondent states that it is easy to
make social connections in Denmark, but that Danes have a tendency to forget these connections
Page 34 of 67

afterwards. As codes relating to hypothesis one only occurs once in the interview, the secondary
does not provide any significant support for Hypothesis 1 either.
Hypothesis 2
The multiple regressions found significant support for Hypothesis 2. The respondent expresses that
there have been issues regarding adapting to the organizational change that followed moving to
Denmark, as the flat structure is perceived to create a degree of chaos. The respondent expresses
discomfort with the Danish organizational structure. However a preference for the low power
distance and the possibility to speak up is mentioned. In the interview the respondent states a
preference towards Denmark, compared to the home country in the case of organizational culture
(Appendix 21). As coding relating to Hypothesis 2, appears seven times throughout the interview,
and gravitates toward a positive influence of factor one, it can be concluded that the qualitative
study triangulates with the results of the primary study (Section 2.7).
Hypothesis 3
Coding related to Hypothesis 1 occurs twice during the qualitative study. The correlation matrix
shows, that there is one correlation between Hypotheses 3 and 7. The respondent explains how
future career perspective might be influenced by language barriers. However, the respondent does
not express whether this has an influence on the motivation to stay or leave. Thus, it is concluded
that language barriers, in alignment with the result in study 1, do not have a significant influence on
the motivation to stay or leave after graduation.
Hypothesis 4
On the basis of the quantitative data (Section 2.7), the primary study does not indicate a significant
influence of Hypothesis 4 influence on the motivation to stay or leave. The code relating to
Hypothesis 4 appears only once, when the respondent is asked about current employment. Thereby
the secondary study does not provide any significant support for Hypothesis 4. This could be
because the questions are not designed to address if Hypothesis 4 in a manner that has an influence
on the motivation to stay or leave.
Hypothesis 5
The statistical analysis shows no significant support for Factor 2, and thereby no support for
Hypothesis 5 either. However, the qualitative study indicates the opposite. The respondent
emphasizes that it is relatively easy to make social connections in Denmark, and that a professional

Page 35 of 67

network definitely have an influence on the possibility of finding a job after graduation, which is
argued to have an influence on the decision to stay or leave. Furthermore, the respondent believes
that good networks increase the chances of getting a job after graduation, which implies that social
factors have an influence on the decision. The correlation between networks and job odds align with
the factor analysis, where the variables social network, professional network and job odds are
clustered into one factor (Network).
Hypothesis 6
The primary study showed no significant support for Hypothesis 6. The respondent expresses
different opinions on the influence of the Danish societal structure on the motivations, depending on
context. The definition of Hypothesis 6 mentions the welfare system as an influence on decision,
which includes taxes, free education, healthcare and etc. In the interview, the respondent states that
the free education was a reason for choosing Denmark as a study destination, but that it does not
influence the decision to stay or leave after graduation. However, the respondent implies that the
high tax level has an influence on the decision. Since high taxes are an essential part of the societal
structure, the contradictory statements indicate a utilization of the societal structure. Moreover, high
taxation is found to have a significant influence on the motivation to stay, after finishing the free
education in Denmark.
It is argued that Hypothesis 6 can be concluded to have an influence on the motivation to leave, in
contrast to the results of the study 1 (Section 2.2.1).
Hypothesis 7
The respondent considers the perspective of getting a job in Denmark low, as a result of poor
Danish language skills. However, the respondent does not directly say whether this influences the
motivation or not. The respondent expresses no current plans on staying in Denmark after
graduation, which could explain why the carrier perspective in Denmark is not considered. It is
argued that there is no significant amount of data, indicating that Hypothesis 7 is effecting the
decision. Thus, confirming the result from study 1 (Section 2.7).

Page 36 of 67

Table 6

3.3 Discussion
3.3.1 Modification for the conceptual model
The qualitative analysis implies that there have to be made some major modifications to the
conceptual model. The hypotheses that are argued not to have an influence on the research question
have to be removed. Thus, Hypothesis 3, 4 and 7 are removed from the model. Consequently the
conceptual model will consist of Hypothesis 1, 2, 5 and 6, which are all argued to have an influence
on the research question.

4 General discussion and conclusion

4.1 Reliability and Validity
In the following section, issues regarding validity and reliability will be addressed. It is important
for the research findings to be accepted as appropriate; therefore all threats and limitations,
affecting the data, need to be taken into consideration,

Page 37 of 67

Reliability indicates that the research is free from errors and bias on the part of the researcher or the
respondent. Reliability is considered crucial for full understanding of the research problem.
Furthermore, reliability of the research is a prerequisite step for obtaining validity.
In order to ensure reliability of this research, a pilot test was conducted, where possible errors and
misunderstandings were addressed before the research was carried out. Moreover, the accuracy and
consistency of the qualitative research was increased due to choice of interviewer. It is argued that
this approach ensures honesty in the respondents answers and an open flow of information.
Nonetheless, to ensure enhanced validity, mixed methods and data source triangulation are used. By
utilizing multiple types of data, including quantitative and qualitative data sources, the validity and
accuracy of the study is increased, allowing for a more precise interpretation.
Above mentioned pilot test, has a positive influence on the overall validity of the quantitative
research. Applying such test, allows the authors to pre-test the survey and detect possible errors and
misunderstandings that can occur during the actual study. As stated previously, study 1 was
conducted online using a random sampling technique. This approach is argued to have a positive
effect on the overall validity of the data output and ensures unbiased selection of respondents.
It has to be taken into account that the sample size for the quantitative was lower than the
requirement set by MacCallum et. al (1999). Furthermore it was recommended that all
communalities should be greater than 0.6. However, two of the factors have communalities below
0.6, thus the requirement is not fully met and decreases the validity of the data. Minor problems
regarding multivariate linearity, multivariate normality and the sample size ratio influence the
overall validity negatively. This means that the correlation between the hypotheses lack a
compelling significant support. A larger sample size could however eliminate the minor assumption
issues in the factor analysis.
Moreover, skewed distributed variables and linearity issues between variables detected during the
analysis, have a negative impact on the overall validity. To compensate for the tendency to
overestimate the fit of the multiple regression model, the adjusted R Square is used to determine to
what degree the model explains the motivation of expatriates to stay in Denmark after graduation.
Using such adjustment will to some degree decrease the negative effect on the overall validity of the
data output.

Page 38 of 67

Additionally, the authors have no control over the conditions in which the respondents have
answered the questionnaire; therefore, the internal validity and diligence of the answers cannot be
certain.
Another, aspect that might have a negative impact of the validity of data output, is the limited
sample size of respondents in study 2. The authors interviewed only one respondent, however since
study 1 is the main study of this research, the limited number of respondents in study 2 is
considered as acceptable. Based on the above, it can be argued that the level of the overall research
validity is acceptable.
External validity indicates whether the relationships found during the research can be generalized to
population other than the research respondents or situation other than the one covered by the
research. It is argued that this research has adequate population validity, meaning that the study can
be applicable to a larger population, like all international students in Denmark. Moreover, the
situation validity is also adequate, since the research can be transferred to other settings. It is
assumed that the knowledge gained from this research, can be applied in a different situation, as
most of the used factors can be suitable to use for instance, when researching expats working in
Denmark. However, due to the fact, that the sample size in both studies is considered as not
optimal, the overall reliability and generalizability of the research is limited.

Page 39 of 67

4.2 Final Conclusion and modifications to the conceptual model

Table 7

All in all, study 2 is triangulating the result in study 1 in regards to hypothesis 1-4 and 7, thus
corroborating the results of the two studies. However, the results of the two studies contradict in
regards to hypothesis 5 and 6, as study 2 indicates a support for these hypotheses. In this context it
is important to emphasize that the overall significance and generalizability of this study is limited
(Section 4.1), especially within study 2, as the study was conducted with only one respondent. As
Factor 5 and 6 were not touched upon in the qualitative study, there is no triangulation for these
factors.

Page 40 of 67

Figure 2 Modified Conceptual Framework

Figure 2 shows how the results of both studies and how these have affected the conceptual
framework (section 1.6). As mentioned before, hypothesis 2 found to have a positive effect on the
research question, therefore, this is illustrated with a +. Through the second study it was
concluded that hypothesis 5 has an influence on the research question. However, the study was not
able to conclude whether the influence was positive or negative. This is illustrated by both + and
-. Study 2 further concluded that hypothesis 6 has a positive influence on the research question,
thus illustrated with a + as well. It is recognized that the conclusion of hypothesis 2 and 6 is weak,
due to the fact that the study was conducted with only one respondent, thus decreasing the validity
and generalizability. As mentioned before it is not possible to connect any hypothesis to Factor 5,
however it was found to have a positive influence on the research question, thus also illustrated with
a+.

4.3 Alternative use of mixed method

This research was conducted with a simple mixed method approach, by conducting a quantitative
study followed by a qualitative study. As the statistical data analysis concluded only one hypothesis
to be significant, it suggests that an alternative use of mixed methods might have been beneficial.

Page 41 of 67

As an example an additional study, a small exploratory qualitative study in the beginning, followed
by quantitative study, and another qualitative study could create the possibility of increased fit of
the model in the multiple regression. The first exploratory study could be an unstructured personal
interview to discover factors influencing the respondents decision to stay or leave after graduation.
An introductory study would provide a better understanding of the issue in question. The
knowledge gained from an exploratory study in combination with previous research, would provide
a better basis for conceptual framework development. The final qualitative study would still be
conducted, with the purpose of triangulation, thus attempting to confirm the conclusion from the
main quantitative study. With this alternative approach the study might have been able to confirm a
better fit of the model in alignment with the conceptual framework, due to the extended background
knowledge. Additionally a broader study, including a larger sample size would have improved the
validity of the research. More respondents would create a possibility to gain significant results in
accordance with the conceptual framework.

Page 42 of 67

Bibliography
Blecher, G. (2013, March 06th). Denmarks Welfare State Works. Retrieved December 1st, 2014,
from The New York Times: http://www.nytimes.com/roomfordebate/2013/05/05/denmarkswork-life-balance/denmarks-welfare-state-works
Creswell, J. W. (2014). Research Design. Qualitative, Quantitative and Mixed Methods Approach.
SAGE Publications.
Dansk Industri. (2010). Fremtiden kalder - uddanner vi nok? Dansk Industri.
Dansk Industri. (2012). Internationale studerende er guld vrd for Danmark. Dansk Industri.
Deloitte. (2008). International Rekruttering - Lsningskatalog. Deloitte.
Fink, A. (2006). How to Conduct Surveys: A Step-by-Step Guide, 3rd ed. Thousand Oaks, CA: Sage
Publications.
Gaard, J. J. (2009). Analyse af Internationale Studerendes Beskftigelsessituation efter endt
Uddannelse i Danmark - Sammenfatning og forslag til initiativer. Forum for Business
Education.
Hansen, M. L. (2010, November 29th). Danmark har 2. hjeste lnninger blandt 33 lande.
Retrieved December 2nd, 2014, from
http://www.cepos.dk/fileadmin/user_upload/dokumenter/Oktober_2010/Danmark_har_2_ho
ejeste_loenninger_blandt_33_lande.pdf
Hofstede, G. (2014a). Organisational Culture & Change Management. Retrieved October 06th,
2014, from The Hofstede Centre: http://geert-hofstede.com/organisational-culture.html
Hofstede, G. (2014b). Natianal Cultural Dimentions. Retrieved October 06th, 2014, from The
Hofstede Centre: http://geert-hofstede.com/national-culture.html
Junge, M., & Jacobsen, R. H. (2009). Analyse af internationale studerende i Danmark
beskftigelsessituationen efter endt uddannelse. Centre For Economic and Business
Research.
MacCallum et. al. (1999). Sample Size in Factor Analysis. Psychological Methods, Vol. 4, No. 1,
pp. 84 - 99.
Maslow, A. H. (1943). A Theory of Human Motivation. Psychological Review, Vol. 50, pp. 370396.
Miles et al. (2014). Qualitative data analysis : an expanded sourcebook, 3rd edition. Thousand
Oaks : SAGE.
Munksgaard, P. G. (2014, April 14th). Thorning: Hjtuddannet arbejdskraft skal skabe vkst.
Retrieved October 06th, 2014, from dr.dk:
http://www.dr.dk/Nyheder/Politik/2014/04/08/125415.htm
Oxford Dictionaries. (2014). Expatriate. Retrieved October 06th, 2014, from OXford Dictionaries:
http://www.oxforddictionaries.com/definition/english/expatriate

Page 43 of 67

Regeringen. (2014b). Lettere adgang til hjkvalificeret arbejdskraft - udspil til en reform af
international rekruttering. Regeringen.
Tabachnick, B. G., & Fidell, L. S. (2012). Using Multivariate Statistics, 6th ed. Pearson Education
Limited.
Uddannelses- og Forskningsministeriet. (2013, March 06th). Internationale studerende styrker
dansk konomi. Retrieved December 7th, 2014, from Pressemeddelelser:
http://ufm.dk/aktuelt/pressemeddelelser/2013/internationale-studerende-styrker-danskokonomi
Zieler, C. (2013, November 07th). Udenlandske studerende koster lille milliard i SU. Retrieved
October 09th, 2014, from Universitetsavisen:
http://universitetsavisen.dk/politik/udenlandske-studerende-koster-lille-halv-milliard-i-su

Page 44 of 67

Appendix 1) Test of normality of chosen variables within the factor

analysis

Page 45 of 67

Appendix 2) Test of linearity between the variables

Page 46 of 67

Appendix 3) SMC values between variables (after exclusion of insignificant variables)

Page 47 of 67

Appendix 4) KMO and Bartletts (after exclusion of insignificant

variables)

Page 48 of 67

Appendix 5) MSA (with all variables before exclusion)

Page 49 of 67

Appendix 6) Rotated component matrix with all variables

Appendix 7) Rotated component Matrix without fair_tax

Page 50 of 67

Appendix 8) Anti-image matrix (excluded fair_tax)

Page 51 of 67

Appendix 9) Rotated component matrix (excluded fair_tax and

dk_individual)

Appendix 10) The best fitting factor matrix

Page 52 of 67

Appendix 11) The regression model

Page 53 of 67

Appendix 12) Histogram of the IVs (test of outliers among the IVs)

Page 54 of 67

Appendix 13) Significance level of all six factors in the regression

model

Appendix 14) Histogram of residuals

Page 55 of 67

Appendix 15) Residuals - checking for normality, linearity and

heteroscedasticity

Page 56 of 67

Appendix 16) Linearity of residuals

Appendix 17) Histogram of DV

Page 57 of 67

Appendix 18: Total Variance expected

Appendix 19: Scree plot

Page 58 of 67

Appendix 20) Communalities for the best fitting model

Page 59 of 67

Appendix 21) Transcript of qualitative interview

Codes:
#

Label

Number of Appearances

Hypothesis 1 Color

National Culture

Hypothesis 2 Color

Organizational Culture

Hypothesis 3 Color

Language barriers

Hypothesis 4 Color

Current employment

Hypothesis 5 Color

Social factors

Hypothesis 6 Color

The Danish societal structure

Hypothesis 7 Color

Carrier perspective in Denmark

Number of correlations
Hypothesis 1 Hypothesis 5 1
Hypothesis 3 Hypothesis 7 1

Transcript
Interviewer: lets start with a simple question, why did you choose to come Denmark?
Respondent: Basically I choose to come to Denmark, because I had a friend here, a Romanian one.
And from the ninth grade, my parents started to say ok, in Bucharest, the capital of my countryRomania, the prices for the accommodation and studying and so on will be like here, maybe 100
Euro less. Difference was very small and they were like joking in that period, saying ok you can go
abroad and study and after that I took into consideration and now I am here.

Page 60 of 67

I chose Denmark because the university was free and also I have heard that they are having a good
system for studying and I dont know I just wanted to try a foreign country and north one because
they are wealthy.
Interviewer: And what are your plans after graduation?
Respondent: At the moment I am planning to finish my Bachelor, after that I suppose I will go for
a Masters Degree in Copenhagen. But I am not sure.
Interviewer: What are your thoughts on working in Denmark after graduation?
Respondent: I dont know if I will work here so much because I want to focus on international
marketing and I would like to work with China at most, because it is the best country in the world.
Where, they have trades and a lot of things. I dont know if I will stay here in Denmark or I will go
in China, but at least I would like to focus on that field.
Interviewer: Do you have a job at the moment?
Respondent: Yes
Interviewer: What kind of job do you have?
Respondent: Well I used to work in a sandwich bar, but I stopped because the exams would come,
and also I am working in Dansk supermarket, so it is not something related to my subject.
Interviewer: Can you describe how you perceive Danish organizational structure. How is it at work
between the employees?
Respondent: The main difference between western countries and European countries the country I
am from, is that they are working in groups, very much, and from time to time I feel like an idiot if I
have to do everything in a group. Because I feel like I am not capable of doing that thing and after
half a year by doing this, when somebody tells me to do something on my own. I am like: Why
dont we do it in a group? This is what I can see and most people focus on relationships and how to
be kind to the other ones, they are not focusing on lets obtain something. They want to have a
strong relationship with the others.
Interviewer: What about the formal relationships with your boss and the others. Is it a flat
organization or more hierarchical

Page 61 of 67

Respondent: I havent seen much hierarchy here, but where I work, there is some. But it depends
from shift to shift, from organization to organization. Because here, especially at school I havent
seen hierarchy that much. Ok he is a director but he is a normal one and I havent seen hierarchy
that much.
Interviewer: How do you like that? Does it appeal to you, to work in such company?
Respondent: Well from my point of view, it is nice to work in that kind of company because your
thoughts are listened and even though they are not taking it into consideration, you can speak up
your mind. But, if it is not hierarchy, there will be chaos, and everybody will do anything they want.
I think that they need from time to time, somebody to stay ok stop, I am the director or I am the
manager and it should be like this.
Interviewer: You describe a little bit of organizational culture already but could you describe it a
little bit more? How do you perceive the culture in Danish companies?
Respondent: The culture in the companies? Well I don't know so much about the culture because I
am not working (study relevant) but as far as I know, they are very open minded. They are looking
forward to working with you even if you don't have that much experience and what I have heard
you just have to go out with the boss for one or two drink and after that you can send him the form
for the application for an internship and he could say 'Ok yes, I was out with you, you are very nice,
outgoing, ok I will take it into consideration' and in the end you already know that you will receive
that job.
Interviewer: Is it different than in your home country?
Respondent: Yes, because in my country it depends on the place you are applying, that you have to
learn a lot to have a lot of studies and make a good impression. This is the old organization because
the newest ones are looking forward to be like here in Denmark but the old one is a hierarchical
system. Where you come to an interview, you speak your mind, you give your cv and you are
waiting for the cell phone, which probably won't be never, nobody will call you at least to say you
know we don't need you.
Interviewer: Based on what you said which culture would you prefer?

Page 62 of 67

Respondent: I would prefer here in Denmark from this point of view. Because I like to have a good
relationship with my boss, not to have fear of him.
Interviewer: Would you be able to have a comfortable discussion with your superior?
Respondent: Well here. Because here you can go to your boss, take a beer, stay with him and to
talk about everything. They listen to you, but in my country basically, if you don't have a strong
relationship with your boss, like if it is a big company you cannot reach the boss so quickly, but if it
is a small company, ok you can talk. Have a chat but not much.
Interviewer: So how would that influence your decision on whether to stay or leave Denmark?
Respondent: This is not influencing my decision of staying or going because there are other factors
like taxes. This is one of the factors that influences my reason of staying or not. But from this point
of view of relationships, I would prefer here.
Interviewer: What about the wage level (the salary)? Is it attractive to you?
Respondent: Here the salary is ok, but they are trying to equalize everybody, everybody who has
the same average. Money to be the same level and the taxes are very high. So it is not appealing for
me from this point of view. But still there are a lot of facilities that you are receiving from those
money.
Interviewer: Do you know about any benefits, that you get from paying taxes?
Respondent: I know that the school is for free. The hospitals are free. There was something else,
but I am not sure now. Basically teaching and going to university is free so this is a strong point.
Interviewer: You mentioned that you want to work with the Chinese market. What are your goals
for your carrier? Do you think you can achieve it in Denmark, if you decide to stay? How do you
see your potential of getting your dream job.
Respondent: in Denmark for the moment my goals are to go to a Summer school in Malaysia and
probably to get a semester abroad like in Shanghai. Because in my country for example, I could go
with Erasmus, but there are a lot of requirements that you have to fulfill in order to go. Here I've
heard there are places left, there are a lot of people that don't go, so you can apply. And you have a
chance to get it. This is from my point of view a good I don't know, a positive factor for them.

Page 63 of 67

Interviewer: But what about your job, carrier wise. Because this is more about school.
Respondent: Well I didn't think so much about my job, because now I am at the school field and it
is a first year here but still when it is about a job, I think I will have more opportunities to go from
here to go abroad than in my home country. For example I couldn't have an opportunity to go so
easily to Shanghai. Or I don't know another country, this is from my point of view a good start.
Interviewer: You are entitled to 5 weeks of paid vacation if you are working in Denmark for over a
year, did you know about that?
Respondent: Yes I knew something. I knew about 9 months and after that you can have a vacation
or something like that. Not a year and after that 5 weeks. I know if you are going in your country
during the vacation, during the holidays, you can have something like a paper from your work to
provide you su that you are not working, but you are still receiving su. Even though you didn't
manage to fulfill your hours.
Interview: Is it the same in your country?
Respondent: Well, we don't have su, but we have something like a scholarship if you are learning
very well. Our system of grades is from 0 to 10 so If you have an average of 8, you receiving those
money. It is not so much, it is like 100 euros, and vacation, yes if you are working, you are
receiving it. I don't know how, because I was not interested but your boss is telling, ok from this
period to this period, everybody has vacation. If you need to take one/three days off, you take into
consideration your days during a year.
Interviewer: How does that motivate you to stay in Denmark? Does it influence the fact that you
have paid 5 weeks vacation?
Respondent: No it doesn't motivate me, because in my country it is the same. So it is not such a big
influence.
Interviewer: Do you find it easy to make social connections in Denmark?
Respondent: It depends with whom I am making social connections. Because it is easy to make
social connections with Danes. But they don't really remember you for a long time. It is like, you
say hello today, you are speaking with them and the other day, ok who are you? But it is ok, it is

Page 64 of 67

easy. You can go and speak to everybody, no one is looking weirdly at you. They are trying to help
you as much as they can. But this is like all.
Interviewer: Are your network connections consisted from more foreigners or Danes?
Respondent: More foreigners I think, and especially Romanians, because I am here like 3 months. I
am trying to find people that are speaking the same mother tongue. To try to understand how it is
going here with everything.
Interviewer: How do you make social connections in Denmark?
Respondent: Firstly my classmates. After that my neighbors and from then I met a lot of new
people, and I am trying to be nice, social and I don't know to make friends.
Interviewer: Is your social network the same as your professional one is it consist of the same
people
Respondent: Well not really, but its almost the same, professional is one or two bosses I am just
seeing them once or twice a week not much we dont talk too much it like ok this is what you have
to do for today and this is like all. If I have questions I can go and they will help me definitely
otherwise no we dont go our things like that.
Interviewer: Do think that this professional network will help you in the future to find a job after
graduation?
Respondent: definitely yes. for example the teachers here, some of them are ok I dont really care
how many do you get 12 or I dont know 4 or 2 I want to know how you are and I want to see how
you are working or if its worth you will receive probably a job but the thing I want for you is to
have a job after a graduation.
Interviewer: How do you see your potential? Of getting a great job after a graduation?
Respondent: Slightly little because I dont know Danish, this would be a weakness, but if I will
study Danish like level 3 it will be ok like medium one I know they can, I can have a starting salary
like 20 thousand.

Page 65 of 67

Interviewer: You mentioned that you like to travel and you said that you are willing to travel as a
part of your studies that was one of my questions too. But if you plan to leave Denmark after
graduation is that decision affected of your need to go to other countries to explore the world?
Respondent: Yes, probably this is true, because I want to see the world as much as I can, and I am
taking advantage of my job. This is why I chose international marketing cause I want to travel the
world and even though I wont have time to visit every city as much as I would like, still I will be
there I will see some main points of that city that country and I will leave.
Interviewer: In what other ways is that are important to you is Denmark different than your home
country?
Respondent: Firstly I will say the nature, cause they have a lot of parks and grass and trees and
they dont have blocks, building blocks, they have just houses, the city is wider than my. They dont
have so many animals, I havent seen pets here, like people with dogs going out and stuff like that,
and children I havent seen playing a lot with each other. but on the other side, I have seen a lot of
people running, even though that outside was a shitty weather and cycling a lot and they are nice
people, I havent seen a lot of shops, like if you need to go to buy something like 12 o'clock in the
evening you wont find anything open just the gas station the prices are very expensive.
Interviewer: What are some other factors that would influence your decision whether to stay or
leave in Denmark after graduation, you still have some years left to maybe you will change?
Respondent: a lot of the decision will be probably again school, I would like to do my master as I
said in Copenhagen but if I have any offer to do it in UK or I dont know or Sweden, or Netherlands
I will go for it because as I said I want to visit as many countries as I can and if I am a student I
have a lot of other benefits from this.
Interviewer: Do you have some suggestions for Danish government to change something to
convince people, graduates to stay in Denmark.
Respondent: As they said, to convince people to stay here is to as they promised before the
election, to give su to every student, even though they work or not. And also any other suggestions,
they should provide also low taxes for the beginning of your work. There are some salaries that said
if you stay here, people between 20 and 30 days to get used to something. If they are having 5, 12,
15% taxes, on they get it on the first year of their salary, they will definitely stay for the second one.

Page 66 of 67

Because they will get used to the boss, with the maids. If they like the environment they are
working in, they will definitely stay.
Interviewer: Thank you very much for your time.

Appendix 22: Survey

https://docs.google.com/forms/d/11Ykkds-wMeVjEg_z_6rbJetkAAIFrT6ftXv_MTjxAck/formResponse

Page 67 of 67